Gene Expression Programming(GEP) genes are structurally organized in a head and a tail with special restrictions and require every symbol in tail must be strictly taken from terminal set.This practice is basically adopted by existing GEP for its perspicuous effect and facility to express,but it is not conducive to semantic computing reuse.This paper seeks to break the restriction on tail and searches a novel open tail GEP algorithm.This algorithm can improve the precision of computing by dynamically introducing the excellent individuals generated during program running to the genes of individuals in a group.The results of symbolic regression experiments show that open tail GEP algorithm outperforms mainstream GEP on average precision performance.
国家科技期刊平台
NSTL
维普
万方数据
